A few ASHP / UFH bits of information.

[ProDave]

I thought your question about the store had been answered.

 

Yes if using a tank like this one that uses a plate heat exchanger, then you need a circulating pump.

 

I would much prefer a cylinder like the one Jack has, with a large area heat input coil. No need for a dedicated pump or the controls that go with it.

 

I also share your concerns about servicability of this system. The control board looked every bit as complicated as a PC mother board, and it's a proprietary item not an off the shelf mother board. So if it goes wrong I just know that's going to be a lot of £££ for a replacement, and in 10 or 20 years will you even be able to buy a replacement?

 

That is why I am so keen to put together a system with simple passive pats and standard central heating control parts. Where that's not possible, any missing control parts will be designed by me and if they go wrong I can fix them, or design a replacement if technology has moved on. That's the theory anyway.  I am currently thinking of using an arduino or a Pi as the controller.

 

[joe90]

I too love the idea of a simple system and I am currently trying to plan mine for installation later this year. Like Peter I also read about the floor one degree above what room temp you want but I am planning to run the ASHP during the night on E7 to wake up with a warm house and only turn it on during the day if required because of very cold weather so I might go for 23-25. I suppose if the mixer on the manifold will only go so low but the ASHP will go lower then the lowest temp rules!. A very good thread which I am following with great interest.

[ProDave]

My only "concern" at the moment about putting together your own control system, is setting the temperature set point of heat pump.

 

The only control interconnection to the heat pump was a 2 core cable between the controller and the heat pump. So it must be using some form of serial interface.

 

Why do I get the feeling that finding the information on what that interface is, and the communication structure is going to be difficult, and even harder to find out before you buy a heat pump

 

At the end of the day, all I want to be able to do is turn the heat pump on and off, and select it's heat output temperature. a requirement that should not be difficult, but with a system sold to be used with a packaged control system that might not be so easy to find out.
 

[PeterW]

18 minutes ago, SteamyTea said:

Heating systems should be designed to last at least 30 years, 50 would be better.  There are no mysteries left at this level of thermodynamics, so we know what is needed and how to do it.

 

 

Most of the newer cylinders such as the Telford ones are guaranteed for 25 years and I've seen one or two with a "lifetime" guarantee. 

 

The heat pump cylinders also have a 3.3m2 coil so that is pretty big when you consider it against a PHE which on that cylinder doesn't look that big. 

[jack]

42 minutes ago, ProDave said:

I am not sold on this weather compensation stuff. I would try and turn that off, or set the curve flat. The house heating controls should take care of putting more heat in when it's colder.

 

At the moment, there are no separate "house heating controls".  We don't even have an internal thermostat.   While I do monitor the slab temp in a couple of places, that temperature isn't presently used to control whether and when the ASHP comes on.  

 

The problem with using house heating controls to call for heat is the huge time constant of the insulated slab.  It takes days to respond fully to a significant change in water temp.  By the time the house is cool enough to demand heat, you're 12-24 hours away from the slab temp rising in a useful way.

 

At the moment, it's just running open loop for a few hours a day, churning out 25-29C water depending on external temp.  The temp range was entirely my decision, and it's easy enough to turn the compensation down or off if desired.  Seems to work fine at the moment, even if the slab temp is probably a degree or so higher than it strictly needs to be right now.  Once we have the insulation in upstairs I'll likely turn down the compensation.

 

8 minutes ago, ProDave said:

At the end of the day, all I want to be able to do is turn the heat pump on and off, and select it's heat output temperature. a requirement that should not be difficult, but with a system sold to be used with a packaged control system that might not be so easy to find out.

 

I suspect with most heat pumps you can call for heat easily enough using volt-free switching.  Setting the temperature is more likely to be handled via a proprietary bus (even if it's based on a comms standard like RS485/Modbus).

 

Last comment about these "all-on-one" packages: the cylinders are generally pretty poorly insulated, and it looks to me like adding insulation will often be difficult.  They are also hugely expensive compared to getting one from a more general cylinder manufacturer.  I may be misremembering, but I seem to recall that the Panasonic cylinder would be over twice the price of a generic one, and that's with less insulation and without the ability to specify the location of extra taps for, eg, thermal sensors.

[jack]

21 minutes ago, joe90 said:

I too love the idea of a simple system and I am currently trying to plan mine for installation later this year. Like Peter I also read about the floor one degree above what room temp you want but I am planning to run the ASHP during the night on E7 to wake up with a warm house and only turn it on during the day if required because of very cold weather so I might go for 23-25. I suppose if the mixer on the manifold will only go so low but the ASHP will go lower then the lowest temp rules!. A very good thread which I am following with great interest.

 

When I first started looking into this, I'm sure I found a supplier that did mixers that went below 25C, but I don't recall who they were, sorry.  I thought I'd give it a go with the standard setup first and so far it's been fine.

[TerryE]

7 hours ago, jack said:

The problem with using house heating controls to call for heat is the huge time constant of the insulated slab.  It takes days to respond fully to a significant change in water temp.  By the time the house is cool enough to demand heat, you're 12-24 hours away from the slab temp rising in a useful way.

 

That's the point I was making above.  The time constant of both the slab and the thermal balance of the house (less solar gain) are very large.  Any room stat no matter how accurate used in on-off mode could be unstable from a control perspective, IMO.  If you are going to use the room stat then you need to smooth it to a similar time constant to the slab's -- well within a factor of 2 or so.  

 

As I've said in other posts I want to run our house for the first year or so with a very simplified but instrumented control regime.  My initial approach will be simple, that is to use an electric inline heater to N kWhr into the slab each night during the economy 7 window (shifted to the end of the window) during periods of net heating, where N is my calculated average daily heat deficit for the house, given the daily average outside temp.  I estimate that this will result in a ripple of less than 2°C in the room temp in the colder months.  (I will use the MVHR bypass to dump excess heat in net cooling months.)

 

This might not be perfect but its going to a huge step improvement over our current house, and this will allow me to collect enough data to design and select my final solution.

 

We are still fitting out the house at the moment so it's unheated, but we've got a small domestic dehumidifier (~300W max) running in the loft to dry out the plaster and painting and this is chuffing out enough heat to keep the whole loft at a comfortable temperature. 

[PeterW]

18 minutes ago, TerryE said:

where N is my calculated average daily heat deficit for the house, given the daily average outside temp.

 

Hi Terry

 

So if I have this correct, you will look at the forecast temperature for the next day and then essentially load the slab with enough "extra" heat to leak into the building from the slab during the day ...? 

 

So if you calculate you need 9KWh then you will run the inline water heater for 3 hours on E7 (assuming it's a 3Kw heater at 100% efficient in this scenario) and then this will be released into the building over the day. 

 

How will you control the release of the heat and where it is released as surely some rooms will overheat..? 

 

Or have I got that all wrong ..??!!

[Stones]

23 hours ago, ProDave said:

That is a VERY interesting price list.

 

A useful find, as you say, illustrates just how much some installers are making, and the 'fat' built into the RRP. 

 

19 hours ago, TerryE said:

 

What concerns me with a conventional controller is that the time constants for slab and the master stat (if based on a room temp) are way out of whack.  It will take about 4kWhr of heat to raise the temperature of my slab by 1°C.  So if you put 4kWhr into the slab when at equilibrium, maybe an hour later the slab will be 1°C warmer and dumping about ½kW additional heat into the room.  The time constants and gains of the two systems are far too separated for a simple on-off control to be stable.   I have some ideas on how to do this, but that's the problem: they are just ideas, because I haven't got my system commissioned and collected enough hard data to characterise the system and enable me to establish a decent control policy. 

 

My experience of a 24/7 weather compensated system (which this is) is based on my last house.  It really was a case of trial and error (in terms of experimenting with different heating curves) to set the system such that the house was at a constant and even temperature.  Once set, you left it alone.  I don't intend operating the slab as a responsive heating system (i.e. whack of the heating if you are cold).  It will be on all the time and do what it needs to when it needs to.  We may very well have heating under or overshoots during the experimentation phase, but working from the starting point of a low 20's flow temp, I don't see this as being a massive issue.  

 

9 hours ago, Barney12 said:

 

@Stones were you furnished with the programming software when you bought from the trade co you linked to?

 

I got everything normally supplied with the heat pump package, instruction, installation manual and a disc.  It should all be programmed in, and simply be a case of working through the menus to set the system up.

 

 

[Stones]

9 hours ago, ProDave said:

I also share your concerns about servicability of this system. The control board looked every bit as complicated as a PC mother board, and it's a proprietary item not an off the shelf mother board. So if it goes wrong I just know that's going to be a lot of £££ for a replacement, and in 10 or 20 years will you even be able to buy a replacement?

 

That is why I am so keen to put together a system with simple passive pats and standard central heating control parts. Where that's not possible, any missing control parts will be designed by me and if they go wrong I can fix them, or design a replacement if technology has moved on. That's the theory anyway.  I am currently thinking of using an arduino or a Pi as the controller.

 

One of the reasons I went for the system was to avoid the need to build custom controls.

 

8 hours ago, jack said:

 

At the moment, there are no separate "house heating controls".  We don't even have an internal thermostat.   While I do monitor the slab temp in a couple of places, that temperature isn't presently used to control whether and when the ASHP comes on.  

 

The problem with using house heating controls to call for heat is the huge time constant of the insulated slab.  It takes days to respond fully to a significant change in water temp.  By the time the house is cool enough to demand heat, you're 12-24 hours away from the slab temp rising in a useful way.

 

At the moment, it's just running open loop for a few hours a day, churning out 25-29C water depending on external temp.  The temp range was entirely my decision, and it's easy enough to turn the compensation down or off if desired.  Seems to work fine at the moment, even if the slab temp is probably a degree or so higher than it strictly needs to be right now.  Once we have the insulation in upstairs I'll likely turn down the compensation.

 

 

How do you know / determine how long to run the open loop for?  

 

 

[MikeSharp01]

22 hours ago, TerryE said:

I have some ideas on how to do this, but that's the problem: they are just ideas, because I haven't got my system commissioned and collected enough hard data to characterise the system and enable me to establish a decent control policy. 

Looks like a classic self tuning PID controller would tackle this Terry, there are many implementations for raspberry PI - such as this one. It would need time to learn, some upper and lower limits carefully set and servos (or digital equivalent) on all the controls but then it should just read the stat and go from there.

[MikeSharp01]

2 hours ago, Stones said:

How do you know / determine how long to run the open loop for?

Just enough time not to crack the slab by heating it too much in one go. This is a worry I have been mulling over. Our polished concrete floor needs to heat and cool very slowly if cracking it to be avoided. So in some ways its good that the time constant is sooooo loooong but if I recall my control theory its all about the sample resolution (digital systems) so you can spot the smallest changes that are above the noise or filter out the noise and get even closer. Sampling the slab depends on the coil pattern a bit I guess as well.

 

Until a couple of weeks back we were not having UFH, didn't need an ASHP but its great to hear that the ASHP can feed the slab direct, that makes perfect sense in a number of scenarios and I am going away to model that and see how close to the optimum COP you can get & stay without cracking the concrete and / or overheating the house.

[Alexphd1]

Not any good for direct of a heat pump but if your trying to control the flow temp from a buffer tank/thermal store then this would do the job:-

http://www.ebay.co.uk/itm/MIXER-THERMOSTATIC-ESBE-VTA-572-1-M-20-43-C-HEATING-28924-/142172210291?hash=item211a1fc473:g:eLkAAOxyCTtTdxgs 

 

i was hoping the bin the UFH mixer and pump and have just one pump between the tank and manifold and use the above valve. 

 

UFH pipes go in tomorrow 

 

 

 

Edited November 30, 2016 by Alexphd1

[Alexphd1]

Not any good for direct of a heat pump but if you're trying to control the flow temp from a buffer tank/thermal store then this would do the job:-

http://www.ebay.co.uk/itm/MIXER-THERMOSTATIC-ESBE-VTA-572-1-M-20-43-C-HEATING-28924-/142172210291?hash=item211a1fc473:g:eLkAAOxyCTtTdxgs 

 

I was hoping to bin the UFH mixer/ zone actuators and pump and have just one pump between the tank and manifold and use the above valve controlled by one stat

 

UFH pipes go in tomorrow 

 

 

 

[TerryE]

4 hours ago, PeterW said:

So if I have this correct, you will look at the forecast temperature for the next day and then essentially load the slab with enough "extra" heat to leak into the building from the slab during the day ...?  So if you calculate you need 9KWh then you will run the inline water heater for 3 hours on E7 (assuming it's a 3Kw heater at 100% efficient in this scenario) and then this will be released into the building over the day.   How will you control the release of the heat and where it is released as surely some rooms will overheat..?

 

@PeterWThe decrement delay for my wall profile is a couple of days so I don't need to be predictive; I can just use the last days average as measured or even the average of the last few days, or even stick a thermal probe into my cavity and take the measurement there (and let the stone skin do the integration for me).  Other than that roughly yes as you say.  I have pretty much the whole slab covered in three loops some at 10cm and some at 15cm centres, but these are run as a single zone so the 9kWh would raise the slab by just over 2°C, though this heat would itself take another hour or more to propagate from the UFH loops throughout the slab and be "seen" at floor level.  So by about 09:00, say, the slab would be pumping an extra 1kW/hr into the living space and this would decay over the day as the room temperature slowly rose, and hence the ripple on the temperature that I mentioned.  I don't control the release of the heat; the whole house is a single zone which has huge inertia.  Clearly if the room temperature is trending down then I will need to trim up the overnight boost.

 

The approach is going to be one of suck it and see. I don't have any upstairs heating at all, so the 1st floor rooms might be a few degrees cooler than the living room and kitchen diner, but that's no bad thing.  I think that the main problem might be my son's room in the loft but that is because he has his gaming PC on most of the evening and it and the rest of his electronics chuff out about 300W plus his personal 200+W of course, plus additional boosts if he has mates around, so it might get a bit warm up there !!    

 

3 hours ago, Stones said:

My experience of a 24/7 weather compensated system (which this is) is based on my last house.  It really was a case of trial and error (in terms of experimenting with different heating curves) to set the system such that the house was at a constant and even temperature.  Once set, you left it alone.  I don't intend operating the slab as a responsive heating system (i.e. whack of the heating if you are cold).  It will be on all the time and do what it needs to when it needs to.  We may very well have heating under or overshoots during the experimentation phase, but working from the starting point of a low 20's flow temp, I don't see this as being a massive issue.

 

@StonesJason, my approach is a KISS baselining for the first year of operation.  Our two approaches share many drivers, but with a different flavour of implementation.  There isn't a single correct approach, and I might well follow your lead after I've got my baseline data.

 

1 hour ago, MikeSharp01 said:

Looks like a classic self tuning PID controller would tackle this Terry...

 

I've lead some embedded control based developments in the early days of my IT career and understand the maths of both linear control.  And yes, I am planning to use an RPi for my main embedded system, running openHAB or Pimatic, with ESP8266s for point collection and control, but that's getting too off-topic.

[Guest Alphonsox]

9 hours ago, Alexphd1 said:

Not any good for direct of a heat pump but if your trying to control the flow temp from a buffer tank/thermal store then this would do the job:-

http://www.ebay.co.uk/itm/MIXER-THERMOSTATIC-ESBE-VTA-572-1-M-20-43-C-HEATING-28924-/142172210291?hash=item211a1fc473:g:eLkAAOxyCTtTdxgs 

 

i was hoping the bin the UFH mixer/ zone actuators and pump and have just one pump between the tank and manifold and use the above valve controlled by one stat

 

UFH pipes go in tomorrow 

 

 

 

 

I found these whilst looking for a low temp mixer (around 40 Euro from Germany), They give a 10-40C adjustment range

http://www.imi-hydronic.com/en/products-solutions/heimeier-thermostatic-control/thermostatic-heads-and-radiator-valves/thermostatic-heads/thermostatic-head-k-with-contact-or-immersion-sensor/

 

At the time @Nickfromwales warned that these are very dependent of the flow/return temperature differentials. I haven't had time to dig any further but will need to before commissioning the UFH.

 

 

[PeterW]

Looks like they have an off the shelf solution for mixing rads and UFH 

 

http://www.imi-hydronic.com/Handlers/FileDownloaderHandler.ashx?path=%2FProductFiles%2FProducts%2Fdocuments%2FCatalogues%2FEnglish_International%2FPDF_low%2FFloor_control_set_EN_lq.pdf

 

 

[jack]

On 11/30/2016 at 19:21, Stones said:

How do you know / determine how long to run the open loop for?  

 

Excellent question, with a simple answer: "I don't"!  

 

If you only ever supply, say, 25C water, then the hottest your rooms can possibly get is 25C.  Warmer than ideal, but not the end of the world.  If that were to happen I could back off the amount of time per day the system is running (or find a mixing valve that allowed for a lower temp, ideally one that's controllable via my home automation system).

 

It goes back to someone's earlier comment about unstable systems - this system can't go out of control because it has a hard limit on its upper temperature that is acceptable (if not ideal). 

 

My recent introduction of a weather compensation curve may possibly challenge this, because the water temp starts increasing from 25C when its 8C outside, until it reaches a maximum temperature of 28/29C (can't remember exactly) at 0C.  Once the insulation is back in the roof, I'll be turning this down a bit, because we really don't need the slab to be sitting at 23-23.5C as it is at the moment.  That said, it's presently -3C outside at 9:30am, and has been even colder the last couple of nights, so it's lovely coming back into the house at the moment!

 

On 11/30/2016 at 21:44, MikeSharp01 said:

Just enough time not to crack the slab by heating it too much in one go. This is a worry I have been mulling over. Our polished concrete floor needs to heat and cool very slowly if cracking it to be avoided. So in some ways its good that the time constant is sooooo loooong but if I recall my control theory its all about the sample resolution (digital systems) so you can spot the smallest changes that are above the noise or filter out the noise and get even closer. Sampling the slab depends on the coil pattern a bit I guess as well.

    

In case it helps, when we turned on our heating a few weeks ago, the slab was at 18C.  It took just under 12 hrs to raise that by 0.5C, which was the fastest rate at which the temperature increased.  The maximum temperature change is therefore pretty-well bang on 1C per 24hrs, with 25C water (no weather compensation)

 

Once the system is up and running each heating season, and ignoring weather compensation, the temperature differential will generally be lower, so the rate of temperature change should be lower too.  What rate of change did your concrete people recommend?  If it's okay for "ordinary" UFH systems, surely it should be fine for a well insulated floor which should have far slower temperature fluctuations with lower temperature differentials?

[Nickfromwales]

5 hours ago, Alphonsox said:

 

I found these whilst looking for a low temp mixer (around 40 Euro from Germany), They give a 10-40C adjustment range

http://www.imi-hydronic.com/en/products-solutions/heimeier-thermostatic-control/thermostatic-heads-and-radiator-valves/thermostatic-heads/thermostatic-head-k-with-contact-or-immersion-sensor/

 

At the time @Nickfromwales warned that these are very dependent of the flow/return temperature differentials. I haven't had time to dig any further but will need to before commissioning the UFH.

 

 

Iirc, @JSHarris has a wunda manifold which utilises a thermo / hydraulic head that requires no such differential as it's not a 'typical' blending valve. 

That uses a hydraulic probe which sits in ( again iirc ) the flow rail of the manifold and regulates the flow temp VERY accurately.  ( his blog gives specifics ). 

[Guest Alphonsox]

I've got the same manifold. I think Jeremy was managing to run at a minimum temp of 22C. Ideally I would like this temperature to be a little lower (18-20C)

[TerryE]

I've been thinking about this issue of the buffer tank.  Jeremy uses one with a precision mix-down as Nick and Neil have described above.  Jason and Dave don't; Jack has one but doesn't use it at the moment.  OK, Jeremy's approach clearly works, but I've really got to ask myself:

 

If you are using the buffer tank for heating only, then why run one at 35°C, say, when you have a humongous concrete buffer at 20°C ?

 

OK, the buffer tank approach is essential if you are using a Gas boiler or equivalent with can't modulate below 10kW, say, but if your heat source is an ASHP which can modulate down to ~ 1kW, then what does it does this complexity really add apart from another tank in your equipment room and a source of heat loss into a small room?

 

The discussion might get a bit nerdy so maybe this is better addressed as a Boffin's corner topic.  Give me a few days and I'll put together a decent strawman to explain my thinking.

 

[joe90]

Why should this be moved to boffins corner, I am very interested in this thread as it WILL effect the design of my heating system. Yes I understand that I may not understand all the technicalities but I will understand the outcomes. I do understand that some may disagree with others, that's life. Keep it coming chaps.

[jack]

1 hour ago, TerryE said:

I've been thinking about this issue of the buffer tank... Jack has one but doesn't use it at the moment. 

 

Nope, I don't have one and have never planned one. The only tank I have is a 250L UVC for DHW. 

[ProDave]

i think my system "design" will be a lot dependent on the actual ASHP I end up with, and at the moment that is an unknown.

 

Because until I have the heat pump, and get to know the ins and outs of how you can control it (and that seems very hard to find out from the blurb pre purchase) I won't know how low a flow temperature it can be set to, and how accurately it will maintain that set flow temperature.

 

If it can indeed be set to say 22 degrees and it will maintain that to say within half a degree, then there will be no need for a regulating UFH manifold and no need for a buffer tank.  But if it can't go that low, or it's control is not very precise then a regulating manifold will be required.

 

What clearly IS needed, for those who are or have set up their own systems, rather than bought a packages system, is just how you can control a particular ASHP and what it can, and more importantly can not do.

 

So what's needed is a technical thread for each make and model of ASHP that forum members have experience of which might help others choose which one to buy to best suit their needs. 

 

[jack]

I had grand plans to bring control of the ASHP into the home automation system, but I suspect that would cost several hundred quid due to the cost of the interfaces that are required for both the ASHP and the home automation system.

 

In the end I just went with the bog standard controller that came with the unit. The instruction manual (which is available online) told me in advance everything I needed to know about connections, programming and info like temperature ranges.

 

Have you narrowed down the likely unit(s) yet? I must say I found Panasonic's tech help line very helpful (we went with an Aquarea)